Fresh install - roamers choppy
 

Fresh install of EQEmu + PEQ database. Roamers such as "a rodent" in cities or.. anywhere really. Sometimes path fine, but sometimes "warp" every few steps. Instead of walking, they appear to be standing, then warp 2-3 feet in the direction they are supposed to be walking.

Then sometimes they start walking normally... then start warping again.

Any idea?

Thanks.

Hey, I dunno if you're still experiencing this issue, but I've noticed this behavior as well after updating my 4 months-old source with the latest updates from git.

After having a look at this particular commit https://github.com/EQEmu/Server/comm...5184bdfb69cb2d, it looks like the call to SendPosition has been moved away from where it was previously.

I tried reverting that change and roamers/pathing behavior went back to normal.

I'm pretty new to the source & codebase so this fix may have unforseen consequences though :p

It would be good if we could have a seasoned dev to comment on this :P

That was in effort to fix the issue, that didn't bring the issue about.

If people are still experiencing warping on the latest, it will need to be looked into more

Totally experiencing it... examples are goblins in runnyeye, rodents in North Qeynos...

Latest code?

Right, sorry Akkadius!

I'm still experiencing it as well, pretty much any NPC assigned to a grid, from what I've seen on my box. As long as they walk a straight line everything's OK, but as soon as they reach their target waypoint and start their journey to the next one, they get choppy and kinda "warp" until they can walk a straight line again.

edit: haven't updated the code for a couple of days though.

Yes, git pull'ed about 8 hours ago. Right now git pull says "Up to date".

#zoned into runnyeye, first goblin peon to encounter is choppy.

And yes, I compiled and relaunched after pulling ;)

I have one of the other devs 'Haynar' looking at this as I am not at home to be doing EQEmu stuff to test/debug.

He should be able to update after he looks at it.

Thanks for the report

I actually reverted back a few functions in mob_ai.cpp back to some from 2015 and my pathing is perfect now.

And what does your diff show with your code compared to the latest in master?

I can share, but I doubt it would help. My source has been changed a ton since the last year. I am far from current, plus a lot of custom changes.

I haven't had a chance to really look at the problem, I can't imagine its all that difficult to resolve, we've just all been busy.

The time period when the got all messed up was back when NPC::AI_SetupNextWaypoint() was added as it's own routine.

But here is a patch, probably only would work with my own source, but here it is anyway.
Never used this service, lets see if it works...

http://pasted.co/27ab2a6c

Finally tested my commit:

https://github.com/EQEmu/Server/comm...5184bdfb69cb2d

The way it was before my change, the mob warps around constantly, right now it is consistent. If people are still seeing it in latest commit I really would like to know what mob pathing they are seeing this on.

This is before my commit, which has the symptoms described:

https://dl.dropboxusercontent.com/u/...pre_change.mp4

This is after my commit:

https://dl.dropboxusercontent.com/u/...ost_change.mp4

This is runspeed 10 on the NPC for sake of speed testing, the NPC is warping before I made the change.

When normal runspeed, the NPC makes normal transitions as well.

This is what is effective in the main code. So I'm not sure if what people are reporting is actually in the most recent code.

Looks like I've been able to reproduce it at runspeed 2

https://dl.dropboxusercontent.com/u/...ue_example.mp4

Ok, resolved the issue:

https://github.com/EQEmu/Server/comm...f4264a2cadeedc

https://dl.dropboxusercontent.com/u/...e_resolved.mp4

Pull latest binaries and you should be good to go.

Praise Akkadius, the Legendary Smasher of Bugs!

Thanks Akkadius, worked well!